"Get stuff done every day. If the task is too large, break it down and explain what you're doing. But get something done every day.
It's too easy for us to say "I'm still working on associating the widgets to the watzits". But with this level of information the customer only hears " - still - ".
This is awesome..
Development: "You want answers?"
Marketing: "I think we are entitled to them!"
Development: "You want answers?!"
Marketing: "I want the truth!"
Development: "You can't handle the truth!!!Son, we live in a world that requires software. And that software must be built by people with elite skills. Who's going to build it? You, Mr. Marketing? You, Mr. Sales? You, Mr. Finance? You, Mr. Human Resources? I don’t think so.
Agile Chronicles: A Few Good Managers
technorati tags:agile, funny, development
Every assertion should be thought from the standpoint of
Translation: assertTrue should always, always have a message.
Consider the following:
assertTrue(mv.getViewName() .startsWith(myController.getSuccessView()));
This will only return "assertion failed". Which is great, but how do we know what happened? If this is buried on one of the lunt automated remote builds, how am I supposed to know what is going wrong? Which is the expected? What actually happened?
A much better version of the same looks like this:
assertTrue( "Was expecting something like " + myController.getSuccessView() + " but was " + mv.getViewName(), mv.getViewName() .startsWith(myController.getSuccessView()));
Same assertion, but now it tells me more specifically what's going on and I can fix bugs with it. Once I set this on the test, it becomes easy to see what was going on.
I recently had a Saturday with some other Bay Area developers where we did a lot of thinking about testing, so expect some more advice in the future as I collect my notes.
Remember the goal of unit tests is to "find bugs" (thanks Harry!). An assertion without an associated message merely notifies you that a bug occurred but doesn't actually "find it". As you are writing your unit tests, make sure you find it as well.
technorati tags:testing, junit, assertions, java
Graeme over at Contrado Solutions has put together some Scrum Poker Cards. These are good for Poker Estimation.
technorati tags:scrum, estimation
I've spent the last few months implementing GTD and I'm still getting used to it, although I feel I'm achieving a comfortable rythm. One of the big "ah-hah" moments in this process however is the difference between our "hunting" and our "gathering" brains. The basic premise of dividing your time in "collecting without doing" and then "doing without collecting" is what makes you feel like you're "in the zone" when you're doing stuff and gives you a relaxed feeling when you're done collecting your stuff.
After doing this for a little bit and then having a small crisis on a Scrum project (where requirements for the current sprint kept changing during the sprint). I realized that one of the big sources of mental stress is precisely trying to do both at once. Our brains just don't seem to be able to handle it.
The way I see it, our brains are pretty much still cavemen brains, used to running out and hunting deer for a very long period of time, and then going out and foraging for food for another long period of time. If you get into a "chase" mode when you're foraging, it means you're the one being chased, and it automatically fires our stressors.
technorati tags:gtd
In this article, Ted Bergeron shows you how to use the Validator component of Hibernate Annotations to build and maintain validation logic easily in your Web apps.
Hibernate can meet your validation needs
technorati tags:java, validation
FullDuplex.org provides us with a funny list of things to do to "look busy".
Here’s a few tips I’ve learned in my short time on this earth that have kept me from getting too much accomplished.
fullduplex.org » How to Do Nothing at Work, and Get Away With It
He goes on to explain his "tips".
It left me thinking two things. First, I'm very glad he's not on my team, second, the value of scrum and agile methodologies to get rid of the dead weight. Here's my response to it:
I'm a lead in a web development team, and I manage and assign work for web developers. This is the behavior that gets you fired from my team.
Let's analyze it shall we?
Look Busy/Look Stressed: We're a Scrum team, we meet every day to talk about what we did the day before, what we will do today, and what roadblocks we have in our way. So why aren't you asking for help? I have no use for web devs with such little people skills that they won't even ask other members of their peer group for help. If you can't be on a team, maybe you should go independent and not work on a team.
Speak Quickly/Hide/Break Limb/Make Excuses/Never Leave your office/rearrange furniture,etc: The only thing here that may work is breaking a limb. And even then if that happened to you it would make you a prime candidate for pairing with another developer. All the others, your non-activity would show up on the cvs commit log. In my team we have a great cvs mailer that shows colored differences of what you did, and I check it daily.
Choose a Profession people don't understand: Your manager is probably not a techie, or he's as lazy as you are about checking your work. You're a perfect example why only techies can manage techies, and techie managers need to code at least 20% of their time on the current project so they can keep their skills. Also following the cvsspam list helps understand the issues.
So you see, using Scrum or other Agile methodologies underscores the lack of value of these people and will make it obvious who needs to drop off your team. Either that or they'll shape up via peer pressure, which in the long run will make the ex-lazy team member happier as well.
technorati tags:agile, scrum, lazy, development
As I watch my little one play with his trains (he's still pre-videogames) I can see the essence of the scientific method playing over and over with his hands and his eyes.
Try to put the big train and attach it to the little train. It doesn't work. How about the toy car? Nope, not that either. At the end he has all the magnetic pieces put together properly and then says "choo-choo!" as he pulls the locomotive. Now he's trying to place a small "Jojo Circus" figure as a "passenger" on top of the train. He pulls it a couple of times but it keeps falling, until he realizes that the bendy legs fit perfectly around the oil tanker car. Then he pulls everything apart and tries it again, this time in a different configuration.
Will Wright on Wired:
Just watch a kid with a new videogame. The last thing they do is read the manual. Instead, they pick up the controller and start mashing buttons to see what happens. This isn't a random process; it's the essence of the scientific method. Through trial and error, players build a model of the underlying game based on empirical evidence collected through play. As the players refine this model, they begin to master the game world. It's a rapid cycle of hypothesis, experiment, and analysis.
Think about this when you write your next piece of software - these are the people you are writing software for (after all, most young adults grew up playing videogames). The previous generation read a lot more than they experimented. The new generations try things out first, then collect their experiences and blog about them, then maybe read someone else's experiences to find kindred spirits. So make your software easy to experiment with, and make most everything undoable.
I am big on flow time. Whenever I am placed in a completely new environment I strive to find a way to achieve the maximum possible flow time.
Here's a great description of this with an eye to human interruptions, which cites De Marco's Peopleware and also makes the connection to ESR's Jargon file.
You should not break without good reason the concentration of a
programmer while they are working. If at all possible, communicate via
e-mail or some other way that lets them react when they have the time.
What follows are some justifications as to why this is so
important.
- Cringe from Crossing a Concentrated Coder, by Lars Wirzenius
But note that this doesn't stop with people, it happens with your choice on technology and tools as well. I think one of the reasons scripting languages have taken off the way they have is that, since they require little or no waiting between changing the file and seeing the results onscreen, they optimize the available flow time even with basic tools (Most of my rails development is currently done with vim and a shell window).
Stop for a few minutes, take a look at your day to day activities and ask yourself when your flow time is typically interrupted by your tools, and whether you can do something to eliminate that interruption. Good tools are designed to reduce or eliminate the flow time interruptions. Bad tools get in the way of that flow time.
Note that if you're only learning to use your current set of tools now, you may find interruptions that are just a part of the learning process - those you can write off as learning time, since they will go away once you master your toolset. But if your process or tool forces you to stop your flow constantly, find a way to improve it or use different tools. You'll be glad you did.
A perfect tool or IDE makes the write-build-test cycle feel so seamless that you forget to think about it, maximizing your flow time.
Finally, note that this optimization is only intended for development time. Your deploy process should still be using build script (i.e., no "building from your IDE" to put in qa or production, please!).
technorati tags: flowtime, programming, agile, flow